Comparing single- and two-channel telephone speech enhancement for elderly hearing-impaired listeners

Abstract

Elderly listeners with sensorineural hearing loss have difficulty understanding speech over the telephone. Hearing aids and telephone amplifiers help, but many elderly individuals do not have the access to these devices. One possible solution to the problem is to ‘‘preprocess’’ the speech at the talker before transmitting it to the hearing-impaired listener. The purpose of this research is to compare processing techniques applied at the talker end that can significantly improve the speech understanding of these persons. Intelligibility enhancement of single- and two-channel amplitude compression schemes are compared by objective and subjective means. The articulation index (AI) and the diagnostic rhyme test (DRT) are used as the objective and subjective measures, respectively. In a preliminary study, we used normal listeners with speech that is passed through Moore’s [(1999)] hearing loss model. The hearing loss model simulates the effects of loudness recruitment and reduced frequency selectivity as well as elevated thresholds of hearing. In addition, elderly hearing-impaired subjects were tested for a subset of the processing schemes. We present results comparing relative intelligibility of the various processing schemes. [Work supported by a grant from the Ohio State University College of Social and Behavioral Sciences.]